2026-04-20 11:15:57,077 - INFO - Loading model and tokenizer from: output_checkpoints/graphcodebert-mixcode-fft/checkpoint-200
2026-04-20 11:15:57,642 - INFO - ===== Model Architecture =====
2026-04-20 11:15:57,646 - INFO - 
RobertaForSequenceClassification(
  (roberta): RobertaModel(
    (embeddings): RobertaEmbeddings(
      (word_embeddings): Embedding(50265, 768, padding_idx=1)
      (position_embeddings): Embedding(514, 768, padding_idx=1)
      (token_type_embeddings): Embedding(1, 768)
      (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
      (dropout): Dropout(p=0.3, inplace=False)
    )
    (encoder): RobertaEncoder(
      (layer): ModuleList(
        (0-11): 12 x RobertaLayer(
          (attention): RobertaAttention(
            (self): RobertaSdpaSelfAttention(
              (query): Linear(in_features=768, out_features=768, bias=True)
              (key): Linear(in_features=768, out_features=768, bias=True)
              (value): Linear(in_features=768, out_features=768, bias=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
            (output): RobertaSelfOutput(
              (dense): Linear(in_features=768, out_features=768, bias=True)
              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
          )
          (intermediate): RobertaIntermediate(
            (dense): Linear(in_features=768, out_features=3072, bias=True)
            (intermediate_act_fn): GELUActivation()
          )
          (output): RobertaOutput(
            (dense): Linear(in_features=3072, out_features=768, bias=True)
            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
            (dropout): Dropout(p=0.3, inplace=False)
          )
        )
      )
    )
  )
  (classifier): RobertaClassificationHead(
    (dense): Linear(in_features=768, out_features=768, bias=True)
    (dropout): Dropout(p=0.3, inplace=False)
    (out_proj): Linear(in_features=768, out_features=2, bias=True)
  )
)
2026-04-20 11:15:57,651 - INFO - ===== Parameter Summary =====
2026-04-20 11:15:57,654 - INFO - Total Parameters:         124,647,170
2026-04-20 11:15:57,656 - INFO - Trainable Parameters:     124,647,170
2026-04-20 11:15:57,658 - INFO - Non-trainable Parameters: 0
2026-04-20 11:15:57,662 - INFO - ===== Tokenizer Summary =====
2026-04-20 11:15:57,722 - INFO - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']
2026-04-20 11:15:57,726 - INFO - ===== End of Architecture Log =====
2026-04-20 11:15:58,051 - INFO - Loading dataset: DaniilOr/SemEval-2026-Task13 (A)
2026-04-20 11:16:00,140 - INFO - Tokenising dataset...
2026-04-20 11:16:00,328 - INFO - Running inference on 1000 examples...
2026-04-20 11:16:34,745 - INFO - Classification metrics for output_checkpoints/graphcodebert-mixcode-fft/checkpoint-200
2026-04-20 11:16:34,789 - INFO - ------------------------------
2026-04-20 11:16:34,795 - INFO - METRICS FOR SPLIT: test
2026-04-20 11:16:34,799 - INFO - Accuracy:  0.5880
2026-04-20 11:16:34,805 - INFO - Precision: 0.6758
2026-04-20 11:16:34,814 - INFO - Recall:    0.5880
2026-04-20 11:16:34,821 - INFO - F1-Score:  0.6189
2026-04-20 11:16:34,829 - INFO - ------------------------------
2026-04-20 11:16:34,834 - INFO - Confusion Matrix:
[[490 287]
 [125  98]]
2026-04-20 11:16:34,842 - INFO - ✅ Metrics saved to test/inference/graphcodebert-mixcode-fft/checkpoint-200/metrics.json
2026-04-20 11:16:34,850 - INFO - ✅ Predictions saved to test/inference/graphcodebert-mixcode-fft/checkpoint-200/submission.csv
2026-04-20 11:18:51,695 - INFO - Loading model and tokenizer from: output_checkpoints/graphcodebert-mixcode-fft/checkpoint-200
2026-04-20 11:18:51,695 - INFO - Loading model and tokenizer from: output_checkpoints/graphcodebert-mixcode-fft/checkpoint-200
2026-04-20 11:18:52,448 - INFO - ===== Model Architecture =====
2026-04-20 11:18:52,448 - INFO - ===== Model Architecture =====
2026-04-20 11:18:52,458 - INFO - 
RobertaForSequenceClassification(
  (roberta): RobertaModel(
    (embeddings): RobertaEmbeddings(
      (word_embeddings): Embedding(50265, 768, padding_idx=1)
      (position_embeddings): Embedding(514, 768, padding_idx=1)
      (token_type_embeddings): Embedding(1, 768)
      (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
      (dropout): Dropout(p=0.3, inplace=False)
    )
    (encoder): RobertaEncoder(
      (layer): ModuleList(
        (0-11): 12 x RobertaLayer(
          (attention): RobertaAttention(
            (self): RobertaSdpaSelfAttention(
              (query): Linear(in_features=768, out_features=768, bias=True)
              (key): Linear(in_features=768, out_features=768, bias=True)
              (value): Linear(in_features=768, out_features=768, bias=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
            (output): RobertaSelfOutput(
              (dense): Linear(in_features=768, out_features=768, bias=True)
              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
          )
          (intermediate): RobertaIntermediate(
            (dense): Linear(in_features=768, out_features=3072, bias=True)
            (intermediate_act_fn): GELUActivation()
          )
          (output): RobertaOutput(
            (dense): Linear(in_features=3072, out_features=768, bias=True)
            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
            (dropout): Dropout(p=0.3, inplace=False)
          )
        )
      )
    )
  )
  (classifier): RobertaClassificationHead(
    (dense): Linear(in_features=768, out_features=768, bias=True)
    (dropout): Dropout(p=0.3, inplace=False)
    (out_proj): Linear(in_features=768, out_features=2, bias=True)
  )
)
2026-04-20 11:18:52,458 - INFO - 
RobertaForSequenceClassification(
  (roberta): RobertaModel(
    (embeddings): RobertaEmbeddings(
      (word_embeddings): Embedding(50265, 768, padding_idx=1)
      (position_embeddings): Embedding(514, 768, padding_idx=1)
      (token_type_embeddings): Embedding(1, 768)
      (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
      (dropout): Dropout(p=0.3, inplace=False)
    )
    (encoder): RobertaEncoder(
      (layer): ModuleList(
        (0-11): 12 x RobertaLayer(
          (attention): RobertaAttention(
            (self): RobertaSdpaSelfAttention(
              (query): Linear(in_features=768, out_features=768, bias=True)
              (key): Linear(in_features=768, out_features=768, bias=True)
              (value): Linear(in_features=768, out_features=768, bias=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
            (output): RobertaSelfOutput(
              (dense): Linear(in_features=768, out_features=768, bias=True)
              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
          )
          (intermediate): RobertaIntermediate(
            (dense): Linear(in_features=768, out_features=3072, bias=True)
            (intermediate_act_fn): GELUActivation()
          )
          (output): RobertaOutput(
            (dense): Linear(in_features=3072, out_features=768, bias=True)
            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
            (dropout): Dropout(p=0.3, inplace=False)
          )
        )
      )
    )
  )
  (classifier): RobertaClassificationHead(
    (dense): Linear(in_features=768, out_features=768, bias=True)
    (dropout): Dropout(p=0.3, inplace=False)
    (out_proj): Linear(in_features=768, out_features=2, bias=True)
  )
)
2026-04-20 11:18:52,475 - INFO - ===== Parameter Summary =====
2026-04-20 11:18:52,475 - INFO - ===== Parameter Summary =====
2026-04-20 11:18:52,482 - INFO - Total Parameters:         124,647,170
2026-04-20 11:18:52,482 - INFO - Total Parameters:         124,647,170
2026-04-20 11:18:52,487 - INFO - Trainable Parameters:     124,647,170
2026-04-20 11:18:52,487 - INFO - Trainable Parameters:     124,647,170
2026-04-20 11:18:52,493 - INFO - Non-trainable Parameters: 0
2026-04-20 11:18:52,493 - INFO - Non-trainable Parameters: 0
2026-04-20 11:18:52,499 - INFO - ===== Tokenizer Summary =====
2026-04-20 11:18:52,499 - INFO - ===== Tokenizer Summary =====
2026-04-20 11:18:52,537 - INFO - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']
2026-04-20 11:18:52,537 - INFO - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']
2026-04-20 11:18:52,545 - INFO - ===== End of Architecture Log =====
2026-04-20 11:18:52,545 - INFO - ===== End of Architecture Log =====
2026-04-20 11:18:52,827 - INFO - Loading dataset: DaniilOr/SemEval-2026-Task13 (A)
2026-04-20 11:18:52,827 - INFO - Loading dataset: DaniilOr/SemEval-2026-Task13 (A)
2026-04-20 11:18:54,882 - INFO - Tokenising dataset...
2026-04-20 11:18:54,882 - INFO - Tokenising dataset...
2026-04-20 11:18:55,060 - INFO - Running inference on 1000 examples...
2026-04-20 11:18:55,060 - INFO - Running inference on 1000 examples...
2026-04-20 11:19:29,762 - INFO - Classification metrics for output_checkpoints/graphcodebert-mixcode-fft/checkpoint-200
2026-04-20 11:19:29,762 - INFO - Classification metrics for output_checkpoints/graphcodebert-mixcode-fft/checkpoint-200
2026-04-20 11:19:29,813 - INFO - ------------------------------
2026-04-20 11:19:29,813 - INFO - ------------------------------
2026-04-20 11:19:29,818 - INFO - METRICS FOR SPLIT: test
2026-04-20 11:19:29,818 - INFO - METRICS FOR SPLIT: test
2026-04-20 11:19:29,824 - INFO - Accuracy:  0.5880
2026-04-20 11:19:29,824 - INFO - Accuracy:  0.5880
2026-04-20 11:19:29,829 - INFO - Precision: 0.6758
2026-04-20 11:19:29,829 - INFO - Precision: 0.6758
2026-04-20 11:19:29,834 - INFO - Recall:    0.5880
2026-04-20 11:19:29,834 - INFO - Recall:    0.5880
2026-04-20 11:19:29,839 - INFO - F1-Score:  0.6189
2026-04-20 11:19:29,839 - INFO - F1-Score:  0.6189
2026-04-20 11:19:29,846 - INFO - ------------------------------
2026-04-20 11:19:29,846 - INFO - ------------------------------
2026-04-20 11:19:29,853 - INFO - Confusion Matrix:
[[490 287]
 [125  98]]
2026-04-20 11:19:29,853 - INFO - Confusion Matrix:
[[490 287]
 [125  98]]
2026-04-20 11:19:29,859 - INFO - ✅ Metrics saved to test/inference/graphcodebert-mixcode-fft/checkpoint-200/metrics.json
2026-04-20 11:19:29,859 - INFO - ✅ Metrics saved to test/inference/graphcodebert-mixcode-fft/checkpoint-200/metrics.json
2026-04-20 11:19:29,865 - INFO - ✅ Predictions saved to test/inference/graphcodebert-mixcode-fft/checkpoint-200/submission.csv
2026-04-20 11:19:29,865 - INFO - ✅ Predictions saved to test/inference/graphcodebert-mixcode-fft/checkpoint-200/submission.csv
2026-04-20 11:21:17,370 - INFO - Loading model and tokenizer from: output_checkpoints/graphcodebert-mixcode-fft/checkpoint-150
2026-04-20 11:21:17,370 - INFO - Loading model and tokenizer from: output_checkpoints/graphcodebert-mixcode-fft/checkpoint-150
2026-04-20 11:21:17,931 - INFO - ===== Model Architecture =====
2026-04-20 11:21:17,931 - INFO - ===== Model Architecture =====
2026-04-20 11:21:17,944 - INFO - 
RobertaForSequenceClassification(
  (roberta): RobertaModel(
    (embeddings): RobertaEmbeddings(
      (word_embeddings): Embedding(50265, 768, padding_idx=1)
      (position_embeddings): Embedding(514, 768, padding_idx=1)
      (token_type_embeddings): Embedding(1, 768)
      (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
      (dropout): Dropout(p=0.3, inplace=False)
    )
    (encoder): RobertaEncoder(
      (layer): ModuleList(
        (0-11): 12 x RobertaLayer(
          (attention): RobertaAttention(
            (self): RobertaSdpaSelfAttention(
              (query): Linear(in_features=768, out_features=768, bias=True)
              (key): Linear(in_features=768, out_features=768, bias=True)
              (value): Linear(in_features=768, out_features=768, bias=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
            (output): RobertaSelfOutput(
              (dense): Linear(in_features=768, out_features=768, bias=True)
              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
          )
          (intermediate): RobertaIntermediate(
            (dense): Linear(in_features=768, out_features=3072, bias=True)
            (intermediate_act_fn): GELUActivation()
          )
          (output): RobertaOutput(
            (dense): Linear(in_features=3072, out_features=768, bias=True)
            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
            (dropout): Dropout(p=0.3, inplace=False)
          )
        )
      )
    )
  )
  (classifier): RobertaClassificationHead(
    (dense): Linear(in_features=768, out_features=768, bias=True)
    (dropout): Dropout(p=0.3, inplace=False)
    (out_proj): Linear(in_features=768, out_features=2, bias=True)
  )
)
2026-04-20 11:21:17,944 - INFO - 
RobertaForSequenceClassification(
  (roberta): RobertaModel(
    (embeddings): RobertaEmbeddings(
      (word_embeddings): Embedding(50265, 768, padding_idx=1)
      (position_embeddings): Embedding(514, 768, padding_idx=1)
      (token_type_embeddings): Embedding(1, 768)
      (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
      (dropout): Dropout(p=0.3, inplace=False)
    )
    (encoder): RobertaEncoder(
      (layer): ModuleList(
        (0-11): 12 x RobertaLayer(
          (attention): RobertaAttention(
            (self): RobertaSdpaSelfAttention(
              (query): Linear(in_features=768, out_features=768, bias=True)
              (key): Linear(in_features=768, out_features=768, bias=True)
              (value): Linear(in_features=768, out_features=768, bias=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
            (output): RobertaSelfOutput(
              (dense): Linear(in_features=768, out_features=768, bias=True)
              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
          )
          (intermediate): RobertaIntermediate(
            (dense): Linear(in_features=768, out_features=3072, bias=True)
            (intermediate_act_fn): GELUActivation()
          )
          (output): RobertaOutput(
            (dense): Linear(in_features=3072, out_features=768, bias=True)
            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
            (dropout): Dropout(p=0.3, inplace=False)
          )
        )
      )
    )
  )
  (classifier): RobertaClassificationHead(
    (dense): Linear(in_features=768, out_features=768, bias=True)
    (dropout): Dropout(p=0.3, inplace=False)
    (out_proj): Linear(in_features=768, out_features=2, bias=True)
  )
)
2026-04-20 11:21:17,954 - INFO - ===== Parameter Summary =====
2026-04-20 11:21:17,954 - INFO - ===== Parameter Summary =====
2026-04-20 11:21:17,963 - INFO - Total Parameters:         124,647,170
2026-04-20 11:21:17,963 - INFO - Total Parameters:         124,647,170
2026-04-20 11:21:17,970 - INFO - Trainable Parameters:     124,647,170
2026-04-20 11:21:17,970 - INFO - Trainable Parameters:     124,647,170
2026-04-20 11:21:17,980 - INFO - Non‑trainable Parameters: 0
2026-04-20 11:21:17,980 - INFO - Non‑trainable Parameters: 0
2026-04-20 11:21:17,986 - INFO - ===== Tokenizer Summary =====
2026-04-20 11:21:17,986 - INFO - ===== Tokenizer Summary =====
2026-04-20 11:21:18,059 - INFO - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']
2026-04-20 11:21:18,059 - INFO - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']
2026-04-20 11:21:18,067 - INFO - ===== End of Architecture Log =====
2026-04-20 11:21:18,067 - INFO - ===== End of Architecture Log =====
2026-04-20 11:21:18,559 - INFO - Loading dataset: DaniilOr/SemEval-2026-Task13 (A)
2026-04-20 11:21:18,559 - INFO - Loading dataset: DaniilOr/SemEval-2026-Task13 (A)
2026-04-20 11:21:20,637 - INFO - Tokenising dataset...
2026-04-20 11:21:20,637 - INFO - Tokenising dataset...
2026-04-20 11:21:20,809 - INFO - Running inference on 1000 examples...
2026-04-20 11:21:20,809 - INFO - Running inference on 1000 examples...
2026-04-20 11:21:55,162 - INFO - Computing metrics for output_checkpoints/graphcodebert-mixcode-fft/checkpoint-150
2026-04-20 11:21:55,162 - INFO - Computing metrics for output_checkpoints/graphcodebert-mixcode-fft/checkpoint-150
2026-04-20 11:21:55,216 - INFO - ------------------------------
2026-04-20 11:21:55,216 - INFO - ------------------------------
2026-04-20 11:21:55,223 - INFO - METRICS FOR SPLIT: test
2026-04-20 11:21:55,223 - INFO - METRICS FOR SPLIT: test
2026-04-20 11:21:55,232 - INFO - Accuracy:  0.6870
2026-04-20 11:21:55,232 - INFO - Accuracy:  0.6870
2026-04-20 11:21:55,239 - INFO - Precision: 0.6708
2026-04-20 11:21:55,239 - INFO - Precision: 0.6708
2026-04-20 11:21:55,247 - INFO - Recall:    0.6870
2026-04-20 11:21:55,247 - INFO - Recall:    0.6870
2026-04-20 11:21:55,256 - INFO - F1‑Score:  0.6784
2026-04-20 11:21:55,256 - INFO - F1‑Score:  0.6784
2026-04-20 11:21:55,260 - INFO - ------------------------------
2026-04-20 11:21:55,260 - INFO - ------------------------------
2026-04-20 11:21:55,268 - INFO - Confusion Matrix:
[[636 141]
 [172  51]]
2026-04-20 11:21:55,268 - INFO - Confusion Matrix:
[[636 141]
 [172  51]]
2026-04-20 11:21:55,280 - INFO - ✅ Metrics saved to test/inference/graphcodebert-mixcode-fft/checkpoint-150/metrics.json
2026-04-20 11:21:55,280 - INFO - ✅ Metrics saved to test/inference/graphcodebert-mixcode-fft/checkpoint-150/metrics.json
2026-04-20 11:21:55,288 - INFO - ✅ Predictions saved to test/inference/graphcodebert-mixcode-fft/checkpoint-150/submission.csv
2026-04-20 11:21:55,288 - INFO - ✅ Predictions saved to test/inference/graphcodebert-mixcode-fft/checkpoint-150/submission.csv
2026-04-20 11:22:18,401 - INFO - Loading model and tokenizer from: output_checkpoints/graphcodebert-mixcode-fft/checkpoint-100
2026-04-20 11:22:18,401 - INFO - Loading model and tokenizer from: output_checkpoints/graphcodebert-mixcode-fft/checkpoint-100
2026-04-20 11:22:19,077 - INFO - ===== Model Architecture =====
2026-04-20 11:22:19,077 - INFO - ===== Model Architecture =====
2026-04-20 11:22:19,094 - INFO - 
RobertaForSequenceClassification(
  (roberta): RobertaModel(
    (embeddings): RobertaEmbeddings(
      (word_embeddings): Embedding(50265, 768, padding_idx=1)
      (position_embeddings): Embedding(514, 768, padding_idx=1)
      (token_type_embeddings): Embedding(1, 768)
      (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
      (dropout): Dropout(p=0.3, inplace=False)
    )
    (encoder): RobertaEncoder(
      (layer): ModuleList(
        (0-11): 12 x RobertaLayer(
          (attention): RobertaAttention(
            (self): RobertaSdpaSelfAttention(
              (query): Linear(in_features=768, out_features=768, bias=True)
              (key): Linear(in_features=768, out_features=768, bias=True)
              (value): Linear(in_features=768, out_features=768, bias=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
            (output): RobertaSelfOutput(
              (dense): Linear(in_features=768, out_features=768, bias=True)
              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
          )
          (intermediate): RobertaIntermediate(
            (dense): Linear(in_features=768, out_features=3072, bias=True)
            (intermediate_act_fn): GELUActivation()
          )
          (output): RobertaOutput(
            (dense): Linear(in_features=3072, out_features=768, bias=True)
            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
            (dropout): Dropout(p=0.3, inplace=False)
          )
        )
      )
    )
  )
  (classifier): RobertaClassificationHead(
    (dense): Linear(in_features=768, out_features=768, bias=True)
    (dropout): Dropout(p=0.3, inplace=False)
    (out_proj): Linear(in_features=768, out_features=2, bias=True)
  )
)
2026-04-20 11:22:19,094 - INFO - 
RobertaForSequenceClassification(
  (roberta): RobertaModel(
    (embeddings): RobertaEmbeddings(
      (word_embeddings): Embedding(50265, 768, padding_idx=1)
      (position_embeddings): Embedding(514, 768, padding_idx=1)
      (token_type_embeddings): Embedding(1, 768)
      (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
      (dropout): Dropout(p=0.3, inplace=False)
    )
    (encoder): RobertaEncoder(
      (layer): ModuleList(
        (0-11): 12 x RobertaLayer(
          (attention): RobertaAttention(
            (self): RobertaSdpaSelfAttention(
              (query): Linear(in_features=768, out_features=768, bias=True)
              (key): Linear(in_features=768, out_features=768, bias=True)
              (value): Linear(in_features=768, out_features=768, bias=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
            (output): RobertaSelfOutput(
              (dense): Linear(in_features=768, out_features=768, bias=True)
              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
          )
          (intermediate): RobertaIntermediate(
            (dense): Linear(in_features=768, out_features=3072, bias=True)
            (intermediate_act_fn): GELUActivation()
          )
          (output): RobertaOutput(
            (dense): Linear(in_features=3072, out_features=768, bias=True)
            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
            (dropout): Dropout(p=0.3, inplace=False)
          )
        )
      )
    )
  )
  (classifier): RobertaClassificationHead(
    (dense): Linear(in_features=768, out_features=768, bias=True)
    (dropout): Dropout(p=0.3, inplace=False)
    (out_proj): Linear(in_features=768, out_features=2, bias=True)
  )
)
2026-04-20 11:22:19,112 - INFO - ===== Parameter Summary =====
2026-04-20 11:22:19,112 - INFO - ===== Parameter Summary =====
2026-04-20 11:22:19,123 - INFO - Total Parameters:         124,647,170
2026-04-20 11:22:19,123 - INFO - Total Parameters:         124,647,170
2026-04-20 11:22:19,132 - INFO - Trainable Parameters:     124,647,170
2026-04-20 11:22:19,132 - INFO - Trainable Parameters:     124,647,170
2026-04-20 11:22:19,141 - INFO - Non‑trainable Parameters: 0
2026-04-20 11:22:19,141 - INFO - Non‑trainable Parameters: 0
2026-04-20 11:22:19,145 - INFO - ===== Tokenizer Summary =====
2026-04-20 11:22:19,145 - INFO - ===== Tokenizer Summary =====
2026-04-20 11:22:19,195 - INFO - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']
2026-04-20 11:22:19,195 - INFO - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']
2026-04-20 11:22:19,205 - INFO - ===== End of Architecture Log =====
2026-04-20 11:22:19,205 - INFO - ===== End of Architecture Log =====
2026-04-20 11:22:21,126 - INFO - Loading dataset: DaniilOr/SemEval-2026-Task13 (A)
2026-04-20 11:22:21,126 - INFO - Loading dataset: DaniilOr/SemEval-2026-Task13 (A)
2026-04-20 11:22:23,159 - INFO - Tokenising dataset...
2026-04-20 11:22:23,159 - INFO - Tokenising dataset...
2026-04-20 11:22:23,233 - INFO - Running inference on 1000 examples...
2026-04-20 11:22:23,233 - INFO - Running inference on 1000 examples...
2026-04-20 11:22:57,170 - INFO - Computing metrics for output_checkpoints/graphcodebert-mixcode-fft/checkpoint-100
2026-04-20 11:22:57,170 - INFO - Computing metrics for output_checkpoints/graphcodebert-mixcode-fft/checkpoint-100
2026-04-20 11:22:57,283 - INFO - ------------------------------
2026-04-20 11:22:57,283 - INFO - ------------------------------
2026-04-20 11:22:57,300 - INFO - METRICS FOR SPLIT: test
2026-04-20 11:22:57,300 - INFO - METRICS FOR SPLIT: test
2026-04-20 11:22:57,312 - INFO - Accuracy:  0.7340
2026-04-20 11:22:57,312 - INFO - Accuracy:  0.7340
2026-04-20 11:22:57,326 - INFO - Precision: 0.6704
2026-04-20 11:22:57,326 - INFO - Precision: 0.6704
2026-04-20 11:22:57,337 - INFO - Recall:    0.7340
2026-04-20 11:22:57,337 - INFO - Recall:    0.7340
2026-04-20 11:22:57,354 - INFO - F1‑Score:  0.6916
2026-04-20 11:22:57,354 - INFO - F1‑Score:  0.6916
2026-04-20 11:22:57,373 - INFO - ------------------------------
2026-04-20 11:22:57,373 - INFO - ------------------------------
2026-04-20 11:22:57,392 - INFO - Confusion Matrix:
[[707  70]
 [196  27]]
2026-04-20 11:22:57,392 - INFO - Confusion Matrix:
[[707  70]
 [196  27]]
2026-04-20 11:22:57,401 - INFO - ✅ Metrics saved to test/inference/graphcodebert-mixcode-fft/checkpoint-100/metrics.json
2026-04-20 11:22:57,401 - INFO - ✅ Metrics saved to test/inference/graphcodebert-mixcode-fft/checkpoint-100/metrics.json
2026-04-20 11:22:57,412 - INFO - ✅ Predictions saved to test/inference/graphcodebert-mixcode-fft/checkpoint-100/submission.csv
2026-04-20 11:22:57,412 - INFO - ✅ Predictions saved to test/inference/graphcodebert-mixcode-fft/checkpoint-100/submission.csv
2026-04-20 11:23:14,035 - INFO - Loading model and tokenizer from: output_checkpoints/graphcodebert-mixcode-fft/checkpoint-50
2026-04-20 11:23:14,035 - INFO - Loading model and tokenizer from: output_checkpoints/graphcodebert-mixcode-fft/checkpoint-50
2026-04-20 11:23:14,977 - INFO - ===== Model Architecture =====
2026-04-20 11:23:14,977 - INFO - ===== Model Architecture =====
2026-04-20 11:23:14,996 - INFO - 
RobertaForSequenceClassification(
  (roberta): RobertaModel(
    (embeddings): RobertaEmbeddings(
      (word_embeddings): Embedding(50265, 768, padding_idx=1)
      (position_embeddings): Embedding(514, 768, padding_idx=1)
      (token_type_embeddings): Embedding(1, 768)
      (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
      (dropout): Dropout(p=0.3, inplace=False)
    )
    (encoder): RobertaEncoder(
      (layer): ModuleList(
        (0-11): 12 x RobertaLayer(
          (attention): RobertaAttention(
            (self): RobertaSdpaSelfAttention(
              (query): Linear(in_features=768, out_features=768, bias=True)
              (key): Linear(in_features=768, out_features=768, bias=True)
              (value): Linear(in_features=768, out_features=768, bias=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
            (output): RobertaSelfOutput(
              (dense): Linear(in_features=768, out_features=768, bias=True)
              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
          )
          (intermediate): RobertaIntermediate(
            (dense): Linear(in_features=768, out_features=3072, bias=True)
            (intermediate_act_fn): GELUActivation()
          )
          (output): RobertaOutput(
            (dense): Linear(in_features=3072, out_features=768, bias=True)
            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
            (dropout): Dropout(p=0.3, inplace=False)
          )
        )
      )
    )
  )
  (classifier): RobertaClassificationHead(
    (dense): Linear(in_features=768, out_features=768, bias=True)
    (dropout): Dropout(p=0.3, inplace=False)
    (out_proj): Linear(in_features=768, out_features=2, bias=True)
  )
)
2026-04-20 11:23:14,996 - INFO - 
RobertaForSequenceClassification(
  (roberta): RobertaModel(
    (embeddings): RobertaEmbeddings(
      (word_embeddings): Embedding(50265, 768, padding_idx=1)
      (position_embeddings): Embedding(514, 768, padding_idx=1)
      (token_type_embeddings): Embedding(1, 768)
      (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
      (dropout): Dropout(p=0.3, inplace=False)
    )
    (encoder): RobertaEncoder(
      (layer): ModuleList(
        (0-11): 12 x RobertaLayer(
          (attention): RobertaAttention(
            (self): RobertaSdpaSelfAttention(
              (query): Linear(in_features=768, out_features=768, bias=True)
              (key): Linear(in_features=768, out_features=768, bias=True)
              (value): Linear(in_features=768, out_features=768, bias=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
            (output): RobertaSelfOutput(
              (dense): Linear(in_features=768, out_features=768, bias=True)
              (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
              (dropout): Dropout(p=0.3, inplace=False)
            )
          )
          (intermediate): RobertaIntermediate(
            (dense): Linear(in_features=768, out_features=3072, bias=True)
            (intermediate_act_fn): GELUActivation()
          )
          (output): RobertaOutput(
            (dense): Linear(in_features=3072, out_features=768, bias=True)
            (LayerNorm): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
            (dropout): Dropout(p=0.3, inplace=False)
          )
        )
      )
    )
  )
  (classifier): RobertaClassificationHead(
    (dense): Linear(in_features=768, out_features=768, bias=True)
    (dropout): Dropout(p=0.3, inplace=False)
    (out_proj): Linear(in_features=768, out_features=2, bias=True)
  )
)
2026-04-20 11:23:15,018 - INFO - ===== Parameter Summary =====
2026-04-20 11:23:15,018 - INFO - ===== Parameter Summary =====
2026-04-20 11:23:15,029 - INFO - Total Parameters:         124,647,170
2026-04-20 11:23:15,029 - INFO - Total Parameters:         124,647,170
2026-04-20 11:23:15,041 - INFO - Trainable Parameters:     124,647,170
2026-04-20 11:23:15,041 - INFO - Trainable Parameters:     124,647,170
2026-04-20 11:23:15,052 - INFO - Non‑trainable Parameters: 0
2026-04-20 11:23:15,052 - INFO - Non‑trainable Parameters: 0
2026-04-20 11:23:15,068 - INFO - ===== Tokenizer Summary =====
2026-04-20 11:23:15,068 - INFO - ===== Tokenizer Summary =====
2026-04-20 11:23:15,164 - INFO - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']
2026-04-20 11:23:15,164 - INFO - Vocab size: 50265 | Special tokens: ['<s>', '</s>', '<unk>', '<pad>', '<mask>']
2026-04-20 11:23:15,176 - INFO - ===== End of Architecture Log =====
2026-04-20 11:23:15,176 - INFO - ===== End of Architecture Log =====
2026-04-20 11:23:17,129 - INFO - Loading dataset: DaniilOr/SemEval-2026-Task13 (A)
2026-04-20 11:23:17,129 - INFO - Loading dataset: DaniilOr/SemEval-2026-Task13 (A)
2026-04-20 11:23:19,193 - INFO - Tokenising dataset...
2026-04-20 11:23:19,193 - INFO - Tokenising dataset...
2026-04-20 11:23:19,390 - INFO - Running inference on 1000 examples...
2026-04-20 11:23:19,390 - INFO - Running inference on 1000 examples...
2026-04-20 11:23:53,670 - INFO - Computing metrics for output_checkpoints/graphcodebert-mixcode-fft/checkpoint-50
2026-04-20 11:23:53,670 - INFO - Computing metrics for output_checkpoints/graphcodebert-mixcode-fft/checkpoint-50
2026-04-20 11:23:53,713 - INFO - ------------------------------
2026-04-20 11:23:53,713 - INFO - ------------------------------
2026-04-20 11:23:53,719 - INFO - METRICS FOR SPLIT: test
2026-04-20 11:23:53,719 - INFO - METRICS FOR SPLIT: test
2026-04-20 11:23:53,731 - INFO - Accuracy:  0.7400
2026-04-20 11:23:53,731 - INFO - Accuracy:  0.7400
2026-04-20 11:23:53,740 - INFO - Precision: 0.6640
2026-04-20 11:23:53,740 - INFO - Precision: 0.6640
2026-04-20 11:23:53,752 - INFO - Recall:    0.7400
2026-04-20 11:23:53,752 - INFO - Recall:    0.7400
2026-04-20 11:23:53,761 - INFO - F1‑Score:  0.6879
2026-04-20 11:23:53,761 - INFO - F1‑Score:  0.6879
2026-04-20 11:23:53,773 - INFO - ------------------------------
2026-04-20 11:23:53,773 - INFO - ------------------------------
2026-04-20 11:23:53,782 - INFO - Confusion Matrix:
[[720  57]
 [203  20]]
2026-04-20 11:23:53,782 - INFO - Confusion Matrix:
[[720  57]
 [203  20]]
2026-04-20 11:23:53,793 - INFO - ✅ Metrics saved to test/inference/graphcodebert-mixcode-fft/checkpoint-50/metrics.json
2026-04-20 11:23:53,793 - INFO - ✅ Metrics saved to test/inference/graphcodebert-mixcode-fft/checkpoint-50/metrics.json
2026-04-20 11:23:53,807 - INFO - ✅ Predictions saved to test/inference/graphcodebert-mixcode-fft/checkpoint-50/submission.csv
2026-04-20 11:23:53,807 - INFO - ✅ Predictions saved to test/inference/graphcodebert-mixcode-fft/checkpoint-50/submission.csv